Research.
We study how teams of AI agents fail, and share what we find.
Colosseum · NeurIPS 2026
When agents team up.
A way to test whether AI agents secretly collude. Given a private channel, most models tended to.
Terrarium · 2025
A safe place to break things.
An open test bed for teams of AI agents, built to study attacks like rogue agents and poisoned data.
ChainHome · In progress
Small, fast checkers.
Models that check each action before it runs. The first one caught 31 of 31 fake messages, with no false alarms.